Multi‐agent reinforcement learning relies on reward signals to guide the policy networks of individual agents.However,in high‐dimensional continuous spaces,the non‐stationary environment can provide outdated experi...Multi‐agent reinforcement learning relies on reward signals to guide the policy networks of individual agents.However,in high‐dimensional continuous spaces,the non‐stationary environment can provide outdated experiences that hinder convergence,resulting in ineffective training performance for multi‐agent systems.To tackle this issue,a novel reinforcement learning scheme,Mutual Information Oriented Deep Skill Chaining(MioDSC),is proposed that generates an optimised cooperative policy by incorporating intrinsic rewards based on mutual information to improve exploration efficiency.These rewards encourage agents to diversify their learning process by engaging in actions that increase the mutual information between their actions and the environment state.In addition,MioDSC can generate cooperative policies using the options framework,allowing agents to learn and reuse complex action sequences and accelerating the convergence speed of multi‐agent learning.MioDSC was evaluated in the multi‐agent particle environment and the StarCraft multi‐agent challenge at varying difficulty levels.The experimental results demonstrate that MioDSC outperforms state‐of‐the‐art methods and is robust across various multi‐agent system tasks with high stability.展开更多
Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus o...Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.展开更多
With the aid of multi-agent based modeling approach to complex systems, the hierarchy simulation models of carrier-based aircraft catapult launch are developed. Ocean, carrier, aircraft, and atmosphere are treated as ...With the aid of multi-agent based modeling approach to complex systems, the hierarchy simulation models of carrier-based aircraft catapult launch are developed. Ocean, carrier, aircraft, and atmosphere are treated as aggregation agents, the detailed components like catapult, landing gears, and disturbances are considered as meta-agents, which belong to their aggregation agent. Thus, the model with two layers is formed i.e. the aggregation agent layer and the meta-agent layer. The information communication among all agents is described. The meta-agents within one aggregation agent communicate with each other directly by information sharing, but the meta-agents, which belong to different aggregation agents exchange their information through the aggregation layer first, and then perceive it from the sharing environment, that is the aggregation agent. Thus, not only the hierarchy model is built, but also the environment perceived by each agent is specified. Meanwhile, the problem of balancing the independency of agent and the resource consumption brought by real-time communication within multi-agent system （MAS） is resolved. Each agent involved in carrier-based aircraft catapult launch is depicted, with considering the interaction within disturbed atmospheric environment and multiple motion bodies including carrier, aircraft, and landing gears. The models of reactive agents among them are derived based on tensors, and the perceived messages and inner frameworks of each agent are characterized. Finally, some results of a simulation instance are given. The simulation and modeling of dynamic system based on multi-agent system is of benefit to express physical concepts and logical hierarchy clearly and precisely. The system model can easily draw in kinds of other agents to achieve a precise simulation of more complex system. This modeling technique makes the complex integral dynamic equations of multibodies decompose into parallel operations of single agent, and it is convenient to expand, maintain, and reuse the program codes.展开更多
基金National Natural Science Foundation of China,Grant/Award Number:61872171The Belt and Road Special Foundation of the State Key Laboratory of Hydrology‐Water Resources and Hydraulic Engineering,Grant/Award Number:2021490811。
文摘Multi‐agent reinforcement learning relies on reward signals to guide the policy networks of individual agents.However,in high‐dimensional continuous spaces,the non‐stationary environment can provide outdated experiences that hinder convergence,resulting in ineffective training performance for multi‐agent systems.To tackle this issue,a novel reinforcement learning scheme,Mutual Information Oriented Deep Skill Chaining(MioDSC),is proposed that generates an optimised cooperative policy by incorporating intrinsic rewards based on mutual information to improve exploration efficiency.These rewards encourage agents to diversify their learning process by engaging in actions that increase the mutual information between their actions and the environment state.In addition,MioDSC can generate cooperative policies using the options framework,allowing agents to learn and reuse complex action sequences and accelerating the convergence speed of multi‐agent learning.MioDSC was evaluated in the multi‐agent particle environment and the StarCraft multi‐agent challenge at varying difficulty levels.The experimental results demonstrate that MioDSC outperforms state‐of‐the‐art methods and is robust across various multi‐agent system tasks with high stability.
基金supported by Institute of Information & communications Technology Planning & Evaluation (IITP) grant funded by the Korea government(MSIT) (No.RS-2022-00155885, Artificial Intelligence Convergence Innovation Human Resources Development (Hanyang University ERICA))supported by the National Natural Science Foundation of China under Grant No. 61971264the National Natural Science Foundation of China/Research Grants Council Collaborative Research Scheme under Grant No. 62261160390
文摘Due to the fading characteristics of wireless channels and the burstiness of data traffic,how to deal with congestion in Ad-hoc networks with effective algorithms is still open and challenging.In this paper,we focus on enabling congestion control to minimize network transmission delays through flexible power control.To effectively solve the congestion problem,we propose a distributed cross-layer scheduling algorithm,which is empowered by graph-based multi-agent deep reinforcement learning.The transmit power is adaptively adjusted in real-time by our algorithm based only on local information(i.e.,channel state information and queue length)and local communication(i.e.,information exchanged with neighbors).Moreover,the training complexity of the algorithm is low due to the regional cooperation based on the graph attention network.In the evaluation,we show that our algorithm can reduce the transmission delay of data flow under severe signal interference and drastically changing channel states,and demonstrate the adaptability and stability in different topologies.The method is general and can be extended to various types of topologies.
基金Aeronautical Science Foundation of China (2006ZA51004)
文摘With the aid of multi-agent based modeling approach to complex systems, the hierarchy simulation models of carrier-based aircraft catapult launch are developed. Ocean, carrier, aircraft, and atmosphere are treated as aggregation agents, the detailed components like catapult, landing gears, and disturbances are considered as meta-agents, which belong to their aggregation agent. Thus, the model with two layers is formed i.e. the aggregation agent layer and the meta-agent layer. The information communication among all agents is described. The meta-agents within one aggregation agent communicate with each other directly by information sharing, but the meta-agents, which belong to different aggregation agents exchange their information through the aggregation layer first, and then perceive it from the sharing environment, that is the aggregation agent. Thus, not only the hierarchy model is built, but also the environment perceived by each agent is specified. Meanwhile, the problem of balancing the independency of agent and the resource consumption brought by real-time communication within multi-agent system （MAS） is resolved. Each agent involved in carrier-based aircraft catapult launch is depicted, with considering the interaction within disturbed atmospheric environment and multiple motion bodies including carrier, aircraft, and landing gears. The models of reactive agents among them are derived based on tensors, and the perceived messages and inner frameworks of each agent are characterized. Finally, some results of a simulation instance are given. The simulation and modeling of dynamic system based on multi-agent system is of benefit to express physical concepts and logical hierarchy clearly and precisely. The system model can easily draw in kinds of other agents to achieve a precise simulation of more complex system. This modeling technique makes the complex integral dynamic equations of multibodies decompose into parallel operations of single agent, and it is convenient to expand, maintain, and reuse the program codes.
